Lecture on Relative Risk and Odds Ratio

Overview

• Part of the Health Stats IQ video series
• Topics covered: Relative Risk, Odds Ratio, Case-Control Studies, Equivalence between RR and OR, Examples from research

Key Concepts

Relative Risk (RR)

• Also known as Risk Ratio
• Measures of group comparison for categorical exposure and outcome
• Used commonly in health sciences (e.g., intervention vs. control group)
• Example: BMI > 30 (obese) vs. BMI < 30 (healthy) in relation to stroke development
• Calculation: Risk of exposed (e.g., obese) / Risk of unexposed (e.g., healthy)
  • Interpretative Example: RR = 2 implies obese individuals have twice the risk of stroke compared to healthy individuals

Odds Ratio (OR)

• Calculated by odds of event in exposed group divided by odds in unexposed group
• Used when RR cannot be computed, such as in case-control studies
• Example: BMI study, OR = 2.25
  • Difference from RR: OR considers odds (e.g., 20/80 for stroke in obese group)
• Necessary for case-control studies (e.g., city living and lung cancer risk)

Case-Control Studies

• Not prospective; select cases (e.g., lung cancer patients) and controls
• Example: Studying the link between inner-city living and lung cancer
  • Challenges: Prospective studies need large samples and long timeframes
  • Data could only reliably calculate ORs, not RRs

Equivalence between RR and OR

• For rare diseases, OR approximates RR closely
• Differences between RR and OR increase with more common diseases
• Chart: Shows relationship between RR and OR across different risk levels

Research Examples

Example 1: Kirhan 2019 - Hearing Loss and Cognitive Decline

• Prospective study using RR
• Findings: Increased risk of cognitive decline with hearing loss
  • E.g., Moderate hearing loss: 1.34 times risk compared to no hearing difficulty

Example 2: Wrench 2020 - Ethnicity and COVID-19

• Case-control study using OR
• Findings: Differences in odds of COVID-19 infection among ethnic groups
  • E.g., Black individuals: 1.93 times odds of positive COVID-19 test compared to white individuals

Conclusion

• OR and RR are critical in health statistics
• OR is essential for case-control studies, while RR is used in prospective studies
• Understanding both measures aids research analysis and interpretation in epidemiology and health sciences

• Encourage further exploration and learning through provided links and resources on zedstatistics.com
• Subscribe and share if interested in health sciences and statistics